Nvidia Pascal Announcement

Pretty sure it will back fire if that is what happens, at this point it seems with a more complex PCB, components and more vram the rx 480 is getting around 30% margins from what Raja stated it wouldn't change their current margins, so with a bit less ram and cheaper PCB, smaller chip, cheaper components, I wouldn't be surprised if nV has a decent higher margin range on this card and at 250, it should be decent margins by nV standards. As you said don't understatement their avarice but going to the point of that much would just hurt them.
 
Please take off those tin-foil hats.

What exactly is Nvidia’s GTX 1080 Founders Edition?

Finally, and this is perhaps the most important point, the Founders Edition cards will not be a limited quantity offer; they will be available at launch and throughout the lifespan of the new Pascal cards. Nvidia also plans to sell the cards directly via its website. So despite the name sounding like something you'd only get if you joined a Kickstarter early on, these will actually be long-term commitments. And that changes the playing field quite a bit. It also explains the higher than MSRP pricing.

There's more to this than just making an alternative card, however. Nvidia has other partners, system vendors, who use their GPUs. Many of these will qualify specific hardware to work in their systems, and some of these vendors really like the old reference cards. They use blowers in place of open air coolers, which means they often work better in SLI configurations or small chassis—Falcon Northwest for example told me they won't put a non-blower GPU in their compact Tiki, due to heat concerns. If the reference design is only available for a short time before transitioning all graphics card production to the AIB partners, it can make things more difficult for system integrators. Now they can qualify the Founders Edition and rest easy knowing the cards will be available for purchase for a year or more.

And that's basically the whole story: the Founders Edition will be an Nvidia manufactured card, just like the old "reference" models, only it will continue to be manufactured and sold throughout the life of the GTX 1080/1070 cards. It will carry a price premium, but that appears to be mostly a case of avoiding too much direct competition with their AIB partners.

http://www.pcgamer.com/what-exactly-is-nvidias-gtx-1080-founders-edition
 
Let me rephrase what I said about the prices:
Why would any AIB inititally sell anything below FE-price. Yes, now the prices have dropped in some countries under FE-prices, but initial AIB-cards were all at FE-price or higher.
Not true. At least in GER and for the 1080, AIB models started at 680 (or 660?) EUR compared to the 789 EUR for the FE.
 
Really doubt it. The 250 is for reference, so the custom would start at 300 or near to it.

The only "reference" card is the FE, and that is being provided only by Nvidia. At $299. AIBs "start from $249 MSRP" but who's really pricing these cards at MSRP when the inferior reference design is more expensive. The $249 is just there for marketing bragging rights while everybody is pricing their cards at FE+ levels. Nvidia can brag about a theoretical low price point even though no card on the market is being priced even close to that.
 
NVIDIA-2014-2017-GPU-Roadmap.png


GP100 - April 2016 - Check
GP104 - June 2016 - Check
GP106 - July 2016 - Check

Next up GP107 - September 2016

FYI: This roadmap was from Jan 2016.
 
The only "reference" card is the FE, and that is being provided only by Nvidia. At $299. AIBs "start from $249 MSRP" but who's really pricing these cards at MSRP when the inferior reference design is more expensive. The $249 is just there for marketing bragging rights while everybody is pricing their cards at FE+ levels. Nvidia can brag about a theoretical low price point even though no card on the market is being priced even close to that.

FE is the reference for 300 dollars ;). All the FE thing is just to sell their cards on higher price while the [strike]naive[/strike] press says they are selling them for lower prices.

;)
 
102 confirmed iirc



Same with Samgsung and Intel and Nvidia.

I don't know where you are but not here, you have some but most people that are will to spend money on their hardware know what they are getting, apple consumers are totally different in this regard, most don't know what is inside apple stuff, and if they did, they would go else where, but they buy apple products because "easy to use" which yeah for a noob or a person that doesn't want to learn they are easier to use for the most part.
 
I don't know where you are but not here, you have some but most people that are will to spend money on their hardware know what they are getting, apple consumers are totally different in this regard, most don't know what is inside apple stuff, and if they did, they would go else were, but they buy apple products because "easy to use" which yeah for a noob or a person that doesn't want to learn they are easier to use for the most part.
Ppl were buying 2 year old 4970Ks for 420 dollars(even used ones.....) and giving them 5 stars. and 450+ 6700. I dont want to get too off topic cuz my post may be deleted but I think it was a complete insult for customers what Intel did with the prices and even worst that ppl were happy to suck it.

As with Apple their products are pretty good, overpriced AF but pretty good, although they lost the innovation factor.
 
NVIDIA-2014-2017-GPU-Roadmap.png


GP100 - April 2016 - Check
GP104 - June 2016 - Check
GP106 - July 2016 - Check

Next up GP107 - September 2016

FYI: This roadmap was from Jan 2016.

GP100 - Tesla P100 announced in April 2016..not Geforce Titan.
GP104 - GTX 1080 launched and available in May 2016..not June 2016.

Next up is GP102.

This roadmap is as much fantasy as it is throwing darts on a board with random names, specs and dates.
 
Judging by GP104 though..demand may just be that high!
You're only allowed to make that argument for the RX 480! (According to nowinstock it's completely sold out everywhere in the US.)

GP106 is 192 bit. And it would be unlikely that GP104 and GP106 are pin compatible so any cut down GP104 product would have to use a GP104 PCB (also due to the higher power requirements). I do wonder why they have those "empty" memory slots though.
If you're sure that the GP106 die is only 192, then the only logical conclusion is that it is ball compatible with GP104. I don't see why that would be a problem? It will allow them to reuse the same PCB with a cut down GP104 and 256 bits for a 1060 Ti or something.

Anyone do a die size estimate yet?
And what's the size of the substrate? If it's the same as GP104, that would almost settle the question about being ball compatible...
 
Ppl were buying 2 year old 4970Ks for 420 dollars(even used ones.....) and giving them 5 stars. and 450+ 6700. I dont want to get too off topic cuz my post may be deleted but I think it was a complete insult for customers what Intel did with the prices and even worst that ppl were happy to suck it.

As with Apple their products are pretty good, overpriced AF but pretty good, although they lost the innovation factor.


There is absolutely no competition for Intel that argument you just stated doesn't have an affect when the market is "artificially" controlled by one company. well not really artificially, it AMD's missteps that has created it. AMD "overcharged" when they had the sizable performance lead too.

Well when you have 50% margins or better ya better have solid products, can't sell crap as gold just doesn't work in today's world.
 
Compatibility with programs written for GP100, for developers who are writing code to deploy on DGX-1 and such. The same reason NVIDIA bothers to include FP64 CUDA cores. NV puts just enough hardware to allow execution, but not enough that it's even remotely fast at it, that way these cards don't undermine Tesla sales. Never mind the fact that the vast majority of GP104 cards will never touch FP16x2 code, so spending too much die space would be wasteful for most cards.
I'm still sort of confused... I can't see how it wouldn't be faster to just do this in "software". Even quantizing after every instruction should only be ~2-4x slower and even handling details like specials and denorms I'm sure you could do it faster than 1/64. Why even bother with the hardware at all?

TBH it doesn't really bother me that they don't have fast fp16 on desktop, but they should have been a lot more upfront about it when they launched consumer Pascal. Lots of folks in the games industry are still working under the incorrect assumption that fp16 is supported and faster on NVIDIA.
 
There is absolutely no competition for Intel that argument you just stated doesn't have an affect when the market is "artificially" controlled by one company. well not really artificially, it AMD's missteps that has created it. AMD overcharged when they had the lead too.

Well when you have 50% margins or better ya better have solid products, can't sell crap as gold just doesn't work in today's world.
My point were not about market share, it was about zombie ppl that buys anything that companies tell them to buy, and it happens not just for apple. The main problem in my opinion are ppl in the USA use its credit card or theirs parents and doesnt bother to think if what they are buying is actually a fair deal.
 
Back
Top